"In considering the drug krokodil, two aspects are of importance, its pharmacology and its chemistry. The short half-life, limited high after the impact effect and, in particular the need for frequent administration may narrow the attention of users on the (circular) process of acquiring, preparing and administering the drug, leaving little time for matters other than avoiding withdrawal and chasing high, as reported in several popular magazines (e.g. Shuster, 2011; Walker, 2011). However, when the layers of bootleg chemistry and attribution are peeled off, what’s left is an opioid analogue (or several ones) that, besides the variations in half-life, behaves pharmacologically not very different than heroin or Hanka (Haemmig, 2011). There are various paths to synthesize desomorphine from codeine, but the chemical process most commonly reported to be used by PWID in Russia and Ukraine is very similar to that of home-produced methamphetamine or Vint (Grund, Zábransky, Irwin, & Heimer, 2009; Zábransky, 2007) – a rudimentary version of a simple chemical reduction. The illicit production of krokodil reportedly involves the processing of codeine into the opiate analogue desomorphine (UNODC, 2012; Gahr et al., 2012a, 2012b, 2012c; Skowronek et al., 2012). Desomorphine (Dihydrodesoxymorphine-D or PermonidTM ) is an opiate analogue first synthesized by Small in 1932 (Small, Yuen, & Eilers, 1933). The analgesic effect of desomorphine is about ten times greater than that of morphine (and thus stronger than heroin), whereas its toxicity exceeds that of morphine by about three times (Weill & Weiss, 1951). The drug’s onset is described as very rapid but its action is of short duration, which may lead to rapid physical dependence and frequent administration."

Source

Grund, J. -P. C., et al. "Breaking worse: The emergence of krokodil and excessive injuries among people who inject drugs in Eurasia." International Journal of Drug Policy (2013), http://dx.doi.org/10.1016/j.d…
http://www.ijdp.org/article/S…